The present invention generally relates to a communication network, and particularly to a method for searching for an alternate path in the communication network when a failure occurs in a node provided therein or a link coupling nodes. More particularly, the present invention is concerned with a method in which each node searches for alternate paths by itself by an improved flooding process.
There are known two different types of alternate path searching processes, one being a centralized control process and the other being a decentralized control process. The centralized control process uses a center station provided in a network. The center station supervises the states of lines. When the center station is informed of the occurrence of a failure from a node, the center station starts to search for alternate paths which do not pass through a failure link.
On the other hand, the decentralized control process uses a flooding procedure in which each node searches for one or more alternate paths by itself. In the flooding procedure, a node which detects a failure sends a message having a destination node address to all links extending from the above node without specifying a route. Each node which receives the message determines whether or not itself is indicated by the received message. Nodes which are not indicated by the message, that is, nodes other than the destination node, pass through the message to all links connected to these nodes. The above-mentioned procedure is repeatedly carried out until the message is received by the destination node. The above-mentioned decentralized control process is known as a routing process in a packet communication, and is highly reliable.
However, the centralized control process needs an increasingly larger amount of time to search for an alternate path as the size of the network increases. Further, the centralized control process needs the center station and cannot search for an alternate path when a failure occurs in the center station.
On the other hand, in the conventional decentralized control process, each node searches for an alternate path, assuming that a failure has occurred in a link. That is, even if a failure occurs in a node, the conventional decentralized control process assumes that the failure has occurred in a link. When a failure occurs in a node and the outputting of a signal stops, a node adjacent to the above node detects the interrupting of the signal supply via a link and concludes that a failure has occurred in the link. Thus, the conventional decentralized control process is not capable of determining whether the failure has occurred in a node or a link, and thus does not cope with a failure which occurs in a node.